The late lamented Dark Sky app was a masterpiece of visualisation that showed you “the shape of the weather”. CC-licensed photo by Susanne Nilsson on Flickr.
You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time. Today: will AI chatbots ease social warming, or make it worse?
A selection of 10 links for you. As forecast. I’m @charlesarthur on Twitter. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
TikTok’s future uncertain after contentious Congress hearing • The Washington Post
Cat Zakrzewski and Jeff Stein:
»
At his first Congressional testimony, TikTok CEO Shou Zi Chew arrived armed with a plan to address mounting national security concerns about the video app’s Chinese owner. But he was met with an unusually hostile and unified coalition of lawmakers, who took turns admonishing him as untrustworthy, in a five-hour thrashing that underscored the wildly popular app’s precarious future in the United States
Lawmakers from both parties sought to tie Chew personally to individuals in the Chinese Communist Party, frequently interrupting him and calling him “evasive.” They repeatedly reminded him that he could face criminal penalties for lying to Congress, as he unsuccessfully attempted to convince them that he could safely steward Americans’ data and shield TikTok from foreign manipulation.
“TikTok is a weapon by the Chinese Communist Party to spy on you, manipulate what you see and exploit for future generations,” said Cathy McMorris Rodgers, the Republican chair of the House Energy and Commerce Committee, who said she supported banning the app outright.
Chew’s appearance thrusts TikTok deeper into a geopolitical standoff between two great economic powers, as support for a ban swells among lawmakers, while key constitutional and legal barriers remain.
Senior Biden administration officials do not believe they have the legal authority to ban TikTok without an act of Congress, according to one person with knowledge of internal government discussions.
«
Strange how things come around again. Trump tried to short circuit the process and effectively sell it off to his friends such as Larry Ellison. But getting Congress to act would be quite the landgrab. And it wouldn’t be the same product: who’d imagine TikTok will hand over the secret algorithmic sauce to a US buyer.
unique link to this extract
Think you could land a plane in an emergency? Here’s why you can’t • The Washington Post
Andrea Sachs:
»
Picture this: You are on a flight when you learn that the pilots have fallen ill and can no longer fly the plane. A voice comes over the public address system, asking for a volunteer to help land the aircraft. You have no experience, but you have seen “Airplane!” and “Snakes on a Plane.” Maybe you’ve frittered away hours on Microsoft’s Flight Simulator. You throw off your seat belt and march toward the cockpit, your cape rustling behind you.
Hold on, hero. You might want to return to your seat for this reality check.
“There is a zero% chance of someone pulling that off,” said Patrick Smith, a commercial air pilot and founder of the Ask the Pilot blog. “Do people think they can perform transplant surgery? No. Then why do they think they can land a plane?”
The clinical name for this type of baseless bravado is the Dunning-Kruger effect. It could be used to explain the results of a YouGov poll conducted in January. Out of 20,063 adults surveyed in the United States, nearly a third said they were “somewhat confident” or “very confident” that they could safely land a passenger airplane in an emergency, relying only on the assistance of air traffic control.
Almost half of the men who responded were confident they could do it, compared with 20% of the women.
«
Dunning-Kruger seems heavily biased towards men. There are similar findings for “could you fight off wild animal [X]?” and “Could you win a point against tennis legend Serena Williams?” Enormous, unjustified overconfidence.
unique link to this extract
Donald Trump shares fake AI-created image of himself on Truth Social • Forbes
Matt Novak:
»
Images created with artificial intelligence have flooded social media in recent months, with some people using AI tools like Midjourney to imagine what it would look like for Donald Trump to be arrested. But the former president isn’t opposed to AI-created photos. Trump shared an image of himself on Thursday morning over at Truth Social. And it’s almost certainly fake.
The image, which has been circulating on pro-Trump Twitter since at least Saturday, shows the former president on one knee praying. At first glance it even looks like it could be a real photo. But anyone who looks closer will notice the telltale signs of AI.
For starters, you always want to look at the hands. AI image creation tools have tremendous difficulty with generating realistic hands, and this image is no different. Trump appears to be missing his ring finger on his right hand, at the very least, and his thumbs are grafted on in a jumbled mess that seems to defy basic human anatomy.
Hugging Face has created a tool that lets people upload images to determine the probability a given image was created with artificial intelligence. The Hugging Face tool says this particular image of Trump was created by machines with about 90% confidence.
But you don’t even need fancy online tools or a knowledge of artificial intelligence to determine this image probably isn’t real. The first clue should’ve been that Trump is taking one knee in a pose that would be somewhat difficult for a fit and healthy man who’s 76 years old, let alone Trump. As far as I can tell, Trump has never taken a knee in public.
«
It’s not “almost certainly” fake – it is fake. As he says, the hands are a dead giveaway. There are multiple other points, such as the completely featureless space. Probably not created by Team Trump, but by religious followers excited at the possibilities of these systems.
Expect a lot more of this.
unique link to this extract
Chinese startup invents long-distance kissing machine • Reuters via The Guardian
»
A Chinese start-up has invented a long-distance kissing machine that transmits users’ kiss data collected through motion sensors hidden in silicon lips, which simultaneously move when replaying kisses received.
MUA – named after the sound people commonly make when blowing a kiss – also captures and replays sound and warms up slightly during kissing, making the experience more authentic, said Beijing-based Siweifushe.
Users can even download kissing data submitted via an accompanying app by other users. The invention was inspired by lockdown isolation. At their most severe, China’s lockdowns saw authorities forbid residents to leave their apartments for months on end.
“I was in a relationship back then, but I couldn’t meet my girlfriend due to lockdowns,” said inventor Zhao Jianbo.
Then a student at the Beijing Film Academy, he focused his graduate project on the lack of physical intimacy in video calls. He later set up Siweifushe which released MUA, its first product, on 22 January. The device is priced at 260 yuan ($38).
In the two weeks after its release, the firm sold over 3,000 kissing machines and received about 20,000 orders, he said.
The MUA resembles a mobile stand with colourless pursed lips protruding from the front. To use it, lovers must download an app on to their smartphones and pair their kissing machines. When they kiss the device, it kisses back.
«
…but expect it to become something with people who get obsessed with their personal chatbots.
unique link to this extract
Tech makers must provide repairs for up to 10 years under proposed EU law • Ars Technica
Scharon Harding:
»
Makers of numerous product categories, including TVs, vacuums, smartphones, and tablets, could be required to enable repairs for their products for up to 10 years after purchase, depending on the device type. The European Commission on Wednesday announced a proposal it has adopted that would implement long-term repair requirements on electronics makers if the European Parliament and Council approve it.
The regulation would apply to any devices with repairability requirements in the EU, including vacuum cleaners, washer-dryers, welding equipment, servers, and data-storage devices. The EU is currently hammering out right-to-repair requirements for smartphones and tablets.
Already, the EU requires vendors to repair or replace products within two years of purchase for free if the product is defective. The new regulation would require companies to provide a free repair (instead of replacing the product) if doing so would be the same price or cheaper than replacing it.
Further, the proposed legislation requires vendors to perform repairs for a minimum of five to 10 years, depending on the device type, after purchase. TV makers, for example, would be required to do repairs for at least seven years after purchase, while washing machine and washer-dryer makers would be on the hook for 10 years. The EU is currently mulling proposals requiring smartphone and tablet makers to provide repairs for up to five years under the law proposed on Wednesday.
The regulation wouldn’t require vendors to perform repairs in this time frame if it is “impossible,” such as if the “repair is technically impossible,” the commission explained in a Q&A page.
«
Which I guess means even AirPods could be repaired, since iFixit has a few repair pages for them.
unique link to this extract
Is Jack Dorsey going to get blown up by Hindenburg Research? • The Verge
Elizabeth Lopatto:
»
How’s our favorite Bitcoin maxi Jack Dorsey doing? Well, the short sellers at Hindenburg Research published an absolute barn-burner of a report alleging widespread fraud at his company, Block. Besides that, Hindenburg says Block misled its investors and is engaging in predatory lending practices.
Oh, okay! Block is threatening to sue. Its shares closed down almost 15% on March 23rd, the day the report was released, from the day before.
If you aren’t familiar with Hindenburg Research, they are bad MFs! Like, they wrote a whole report alleging that the electric vehicle company Nikola was “intricate fraud built on dozens of lies over the course of its Founder and Executive Chairman Trevor Milton’s career.” Milton was later convicted of fraud. They also went after Lordstown Motors, saying its executives had made misleading claims about truck preorders. Those allegations appeared to be substantiated by a law firm’s investigation into the preorders — statements about preorders were “in certain respects, inaccurate” was the phrasing. The CEO was forced out, and the Department of Justice decided they wanted to investigate.
This is also how Hindenburg makes money. They’re short sellers — which means they make money by betting a stock will decline in value. After they make their bets, they release their report, which, yes, often makes the stock decline in value! It’s cool that anyone does this much research and reporting on companies; I know very few journalists who can spend two years on a single story like this.
«
But for Hindenburg, of course, they can spend two years on something if they expect the payoff to be substantial enough. And we’re talking about multiple millions here. On the basis, short sellers are a form of journalism: telling the stories that afflict the comfortable. It’s just there tends not to be much in the way of public service there.
unique link to this extract
A eulogy for Dark Sky, a data visualization masterpiece • Nightingale
Srini Kadamati, in the journal of the Data Visualisation Society:
»
On January 1, 2023, Apple sunsetted (pun intended) the Dark Sky mobile app on iOS. Apple purchased the company behind the popular weather application in early 2020, then announced that it would be shutting down the Dark Sky applications (first on Android, then on iOS and web), and finally stated in 2022 that the forecast technology would be integrated into the Apple Weather app with iOS 16.
But Dark Sky was much more than just an API or a set of “forecast technologies.” The design of the Dark Sky mobile application represented a hallmark of information design because the team clearly obsessed over how people would actually use the app on a daily basis.
«
This is a wonderful article, which also serves as a reminder of how much we lost when Apple incorporated it into its Weather app. (Subsequently Apple does seem to have adopted some of the smarter elements of Dark Sky, such as the “shape” of the weather for the day – though not as deftly done as Dark Sky did.)
unique link to this extract
I saw the face of God in a TSMC semiconductor factory • WIRED
Virginia Heffernan:
»
Like a dutiful valet who exists only to make his aristocrat look good, TSMC supplies the brains of various products but never claims credit. The fabs operate offstage and under an invisibility cloak, silently interceding between the flashy product designers and the even flashier makers and marketers. TSMC seems to relish the mystery, but anyone in the business understands that, were TSMC chips to vanish from this earth, every new iPad, iPhone, and Mac would be instantly bricked. TSMC’s simultaneous invisibility and indispensability to the human race is something that Jensen Huang, the CEO of Nvidia, likes to joke about. “Basically, there is air—and TSMC,” he said at Stanford in 2014.
“They call Taiwan the porcupine, right? It’s like, just try to attack. You may just blow the whole island up, but it will be useless to you,” Keith Krach, a former US State Department undersecretary, told me a few weeks before I left for Taiwan. TSMC’s chairman and former CEO, Mark Liu, has put it more concretely: “Nobody can control TSMC by force. If you take by military force, or invasion, you will render TSMC inoperative.” If a totalitarian regime forcibly occupied TSMC, in other words, its kaiser would never get its partner democracies on the phone. The relevant material suppliers, chip designers, software engineers, 5G networks, augmented-reality services, artificial-intelligence operators, and product manufacturers would block their calls. The fabs themselves would be bricked.
«
I’m not going to pretend that this article is short, or doesn’t have longeurs that could easily have been left out. But it’s the first time we know of that a journalist has been inside TSMC.
unique link to this extract
Was this written by a human or AI? ¯_(ツ)_/¯ • Stanford University Human-Centered AI
Prabha Kannan:
»
What are the implications and risks of using AI-generated text, especially in online dating, hospitality, and professional situations, areas where the way we represent ourselves is critically important to how we are perceived?
“Do I want to hire this person? Do I want to date this person? Do I want to stay in this person’s home? These are things that are deeply personal and that we do pretty regularly,” says Jeff Hancock, professor of communication at Stanford School of Humanities and Sciences, founding director of Stanford’s Social Media Lab, and a Stanford Institute for Human-Centered AI faculty member. Hancock and his collaborators set out to explore this problem space by looking at how successful we are at differentiating between human and AI-generated text on OKCupid, AirBNB, and Guru.com.
What Hancock and his team learned was eye-opening: participants in the study could only distinguish between human or AI text with 50-52% accuracy, about the same random chance as a coin flip.
The real concern, according to Hancock, is that we can create AI “that comes across as more human than human, because we can optimize the AI’s language to take advantage of the kind of assumptions that humans have. That’s worrisome because it creates a risk that these machines can pose as more human than us,” with a potential to deceive.
«
More than a “potential”, I’d say. A certainty, more like. The only question is where.
unique link to this extract
Microsoft justifies AI’s ‘usefully wrong’ answers • CNBC
Jonathan Vanian:
»
On Tuesday, Google announced it was bringing AI-powered chat technology to Gmail and Google Docs, letting it help composing emails or documents. On Thursday, Microsoft said that its popular business apps like Word and Excel would soon come bundled with ChatGPT-like technology dubbed Copilot.
But this time, Microsoft is pitching the technology as being “usefully wrong.”
In an online presentation about the new Copilot features, Microsoft executives brought up the software’s tendency to produce inaccurate responses, but pitched that as something that could be useful. As long as people realize that Copilot’s responses could be sloppy with the facts, they can edit the inaccuracies and more quickly send their emails or finish their presentation slides.
For instance, if a person wants to create an email wishing a family member a happy birthday, Copilot can still be helpful even if it presents the wrong birth date. In Microsoft’s view, the mere fact that the tool generated text saved a person some time and is therefore useful. People just need to take extra care and make sure the text doesn’t contain any errors.
Researchers might disagree.
Indeed, some technologists like Noah Giansiracusa and Gary Marcus have voiced concerns that people may place too much trust in modern-day AI, taking to heart advice tools like ChatGPT present when they ask questions about health, finance and other high-stakes topics.
«
Google seems to be rushing here, where Microsoft is usefully positioned – adding chatbot functionality to Office takes it into a much more compelling space.
unique link to this extract
| • Why do social networks drive us a little mad? • Why does angry content seem to dominate what we see? • How much of a role do algorithms play in affecting what we see and do online? • What can we do about it? • Did Facebook have any inkling of what was coming in Myanmar in 2016? Read Social Warming, my latest book, and find answers – and more. |
Errata, corrigenda and ai no corrida: none notified
The Overspill: when there's more that I want to say 